@jakub-o: It seems that there was an regression in kernel 3.9 which introduced this issue for certain users. Which kernel do you use at the moment ?
Do you get a similar dmesg output ?
In kernel 3.10 a patch was applied which brings some improvement in this situation according to a number of testers and myself. Maybe it brings relief to you. If you want and have the knowledge you can test it yourself [1].
Also the Chromium guys proposed a patch similar patch [2] which could bring some improvement too. If you want you can build your own kernel using the patch [3]. For me it brought no improvement, but I would be good to get more test data for it.
@jakub-o: It seems that there was an regression in kernel 3.9 which introduced this issue for certain users. Which kernel do you use at the moment ?
Do you get a similar dmesg output ?
In kernel 3.10 a patch was applied which brings some improvement in this situation according to a number of testers and myself. Maybe it brings relief to you. If you want and have the knowledge you can test it yourself [1].
Also the Chromium guys proposed a patch similar patch [2] which could bring some improvement too. If you want you can build your own kernel using the patch [3]. For me it brought no improvement, but I would be good to get more test data for it.
[1] https:/ /wiki.ubuntu. com/Kernel/ MainlineBuilds# Mainline_ Kernels_ Archive
[2] https:/ /patchwork. kernel. org/patch/ 2156601/
[3] https:/ /wiki.ubuntu. com/Kernel/ BuildYourOwnKer nel